Re: [Mingw-users] Re: Make configuration for Win32.

[Paul D. S. <ps...@gn...>]

%% "Luke Dunstan" <cod...@ho...> writes:

  ld> I don't quite understand what is wrong with _WIN32, but I also
  ld> wouldn't object to __WINDOWS32__, etc. (but I agree with Danny
  ld> that WINDOWS32 is no good for a compiler built-in).

Wait--please remember that WINDOWS32 is _NOT_ a compiler built-in, nor
have I ever suggested it should be.

It's the macro that GNU make uses in its own code when it wants to build
for a Windows platform.

It's you-all who are talking about defining new compiler builtin
macros.  I personally see no real need for it.

  ld> Since only Mingw (and Cygwin I guess) will define __WINDOWS32__
  ld> (AFAIK), checking for this macro will really just be checking for
  ld> Mingw, but if that's what you want to do then __MINGW32__ would be
  ld> a better option. Then again, you probably need special build
  ld> processes for MSVC and other Windows compilers anyway, so
  ld> explicitly defining __WINDOWS32__ for those wouldn't be a problem.

Maybe I'm misunderstanding what you're trying to accomplish in terms of
changes to GNU make.

Right now GNU make has its own macro, WINDOWS32, as I said above.  When
you compile with MSVC etc. you use a predefined config.h which contains
hardcoded results for a Windows platform, and defines this macro.

If the MINGW folks want to allow GNU make to be autoconf-enabled on a
MINGW system, that's fine with me.  But I don't understand the purpose
of changing GNU make's personal macro.


Why not just check the system type in the configure script, and, as I
suggested before, just use AC_DEFINE(WINDOWS32, 1,[...]) if it's a
Windows platform so that macro is defined in config.h?

-- 
-------------------------------------------------------------------------------
 Paul D. Smith <ps...@gn...>          Find some GNU make tips at:
 http://www.gnu.org                      http://make.paulandlesley.org
 "Please remain calm...I may be mad, but I am a professional." --Mad Scientist